DWORKIN’S LIVING WELL AND THE WELL-BEING REVOLUTION

* CHRISTINE JOLLS

INTRODUCTION ...... 641 I. LIVING WELL AND SELF-REPORTED WELL-BEING: DEFINITIONS ...... 643 II. GENUINENESS ...... 645 III. SELF-REPORTED WELL-BEING, INCOME, AND GENUINENESS ...... 650 IV. SELF-REPORTED WELL-BEING, SOCIABILITY, AND GENUINENESS ..... 651 V. LIVING WELL? ...... 653 CONCLUSION ...... 654

Happiness is like the summer, it does not go on radiating. Nothing to be hoped from remembering it, on days when we feel cold. There are sensations which write themselves in indelible characters. Happiness writes white.1

INTRODUCTION Philosophers from Aristotle to Mill to Dworkin have considered the relationship between what it means to “live well” in our own lives (“ethics” in Ronald Dworkin’s Justice for Hedgehogs2) and how we ought to treat others (“morality”).3 Far from any notion that morality operates as a dispiriting constraint on “living well,” Dworkin – like Aristotle – views ethics and morality as deeply complementary. For Aristotle, the state of eudaimonia, or happiness, is “the best, noblest, and most pleasant thing in the world, and these attributes are not severed as in the inscription at Delos – ‘Most noble is that

* Gordon Bradford Tweedy Professor of Law and Organization, Yale Law School. This Essay was prepared for the Boston University School of Law Symposium, Justice for Hedgehogs: A Conference on Ronald Dworkin’s Forthcoming Book, September 25-26, 2009. I am grateful to Bruce Ackerman, Ian Ayres, Richard Fallon, Joseph Fishkin, Dov Fox, Jake Lucchi, Tienmu Ma, Daniel Markovits, John Muller, Robert Post, Aaron Voloj, participants at the Yale Law School Faculty Workshop, and conference participants at Boston University School of Law for extremely helpful comments, and to Eric Sublett for helpful interactions over the dataset on which this Essay draws. 1 HENRY DE MONTHERLANT, COSTALS AND THE HIPPOGRIFF 157 (John Rodker trans., Alfred A. Knopf, Inc. 1940) (1937). 2 RONALD DWORKIN, JUSTICE FOR HEDGEHOGS (forthcoming 2010) (Apr. 17, 2009 manuscript on file with the Boston University Law Review). 3 See id. (manuscript at 8). 641

642 BOSTON UNIVERSITY LAW REVIEW [Vol. 90:641 which is justest, and best is health; But most pleasant it is to win what we love.’”4 Dworkin likewise rejects the inscription at Delos, urging instead that the “truth about living well and being good . . . is not only coherent but mutually supporting.”5 Justice for Hedgehogs seeks “to illustrate as well as defend the unity of at least ethical and moral values.”6 If it is clear enough that “winning what we love” is not sufficient for “living well” for Dworkin, it also seems clear that people’s own perceptions of their satisfaction or well-being are of at least some import on a wide range of philosophical views; as Thomas Scanlon writes, “[A]ny plausible substantive good theory will count agreeable mental states among the things which can make a life better.”7 To take a somewhat narrow example, empirical surveys of people’s perceived well-being consistently find that well-being is highly correlated with the body’s successful deployment of immune responses to viral and other infections8 – and one would think avoiding various forms of infectious illness would make life “better” to at least some degree on most accounts. This Essay shall put into conversation the constructs of “living well” and self-reported well-being by bringing into perhaps surprising – if inevitably limited – contact Dworkin’s ethics and a new methodology that in recent years has revolutionized the empirical measurement of perceived well-being. A research team that I shall call the “Princeton group,” consisting of Daniel Kahneman, Alan Krueger, David Schkade, Norbert Schwarz, and Arthur Stone, has produced an immensely textured – and entirely publicly available – dataset on the nature, level, and correlates of people’s self-reported well- being.9 The Princeton group’s data are rich in many ways but most crucially, for present purposes, in their embrace of diverse and at times conflicting ways of measuring people’s perceived well-being. Strikingly, some of the most interesting conflicts in the data may be mapped, though in only a partial way, onto a central piece of Dworkin’s architecture for understanding the meaning of “living well.” This mapping both deepens an important aspect of the

4 ARISTOTLE, THE NICOMACHEAN ETHICS bk. I, at 17 (J.L. Ackrill & J.O. Urmson eds., David Ross trans., Oxford Univ. Press 1992) (c. 384-322 B.C.E.). 5 DWORKIN, supra note 2 (manuscript at 7). 6 Id.; see also id. (manuscript at 122) (“We want to think that morality connects with human purposes and ambitions . . . , that it is not all constraint and no value.”). 7 Thomas Scanlon, Value, Desire, and Quality of Life, in THE QUALITY OF LIFE 185, 192 (Martha Nussbaum & Amartya Sen eds., 1993). 8 A comprehensive survey of this evidence can be found in Alan B. Krueger, Daniel Kahneman, David Schkade, Norbert Schwartz & Arthur A. Stone, National Time Accounting: The Currency of Life, in MEASURING THE SUBJECTIVE WELL-BEING OF NATIONS 9, 25-26 (Alan B. Krueger ed., 2009). 9 See generally id.; sources cited infra notes 17, 19, 28. The complete dataset is available at http://www.krueger.princeton.edu/subjective.htm (select “Tx DRM Stata Data File”) (last visited Jan. 20, 2010).

2010] “LIVING WELL” AND WELL-BEING 643

Princeton group’s analysis and serves to highlight an intriguing question about Dworkin’s own approach.10 First things first: how does Dworkin define “living well,” and in what ways can we begin to relate his approach to the empirical literature on self-reported well-being?11

I. LIVING WELL AND SELF-REPORTED WELL-BEING: DEFINITIONS In an engaging and thought-provoking parallel, Dworkin analogizes “living well” to artistic creation. Central to both, he suggests, is the performance – of life or of the act of artistic creation – rather than the product – the “completed narrative” of a life or the piece of artwork produced: We value great art most fundamentally not because the art as product enhances our lives but because it embodies a performance, a rising to artistic challenge. We value human lives as these are lived not for the completed narrative, as if fiction would do as well, but because they too embody a performance: rising to the challenge of having a life to lead.12 For Dworkin, “The final value of our lives is adverbial . . . . It [is] the value of the performance . . . .”13 Dworkin’s product-performance distinction is striking from the perspective of the recent empirical literature on self-reported well-being because a central topic in that literature is the disjunction between well-being measures that elicit evaluative views of the overarching “completed narrative”14 of one’s life and well-being measures that focus on views of the actual “performance”15 of life’s tasks and activities.16 Consider the leading examples of each type of measure.

10 See infra text accompanying notes 55-57 for discussion of the question raised about Dworkin’s approach. 11 Like a number of the other contributors to this Symposium, I attempt to focus in on one of the many moving and interconnected pieces in Dworkin’s ambitious book. See, e.g., Richard H. Fallon, Jr., Is Moral Reasoning Conceptual Interpretation?, 90 B.U. L. REV. 535, 535-36 (2010) (“Because of the interconnections among the topics that Dworkin covers and the mutual support that his various arguments afford one another, it is nearly impossible to assess any of his claims without struggling with, and either accepting or rejecting, the overachieving philosophical position that he, as a self-advertised hedgehog, very deliberately presents as ‘one big thing.’ To the greatest possible extent, however, I shall . . . [f]ocus[] . . . on Dworkin’s claim that moral reasoning is conceptual interpretation . . . .” (citation omitted)). My hope is that by focusing in on a particular aspect of Dworkin’s overall edifice – his conception of “living well” – some illumination of this aspect of his work and its implications for alternative approaches will be possible. 12 DWORKIN, supra note 2 (manuscript at 124). 13 Id. 14 Id. 15 Id. 16 See infra text accompanying notes 30-54.

644 BOSTON UNIVERSITY LAW REVIEW [Vol. 90:641

Overarching satisfaction with one’s life has long been measured through questions such as “All things considered, how satisfied are you with your life as a whole . . . ?” (World Values Survey) and “Taken all together, how would you say things are these days? Would you say that you are very happy, pretty happy, or not too happy?” (General Social Survey).17 Data on these sorts of questions exists for dozens of countries over substantial time periods.18 Whatever the limitations (which are many) of this measure of well-being, it seems clearly targeted to a global, overarching life assessment. In sharp contrast, the so-called Day Reconstruction Method (“DRM”) pioneered by the Princeton group in recent years measures well-being by reference to assessment of the performance of particular activities within a life. The DRM asks participants to “revive memories of the previous day by constructing a diary consisting of a sequence of episodes” and then to “describe each episode by answering questions about the situation and about the feelings that they experienced” during that episode.19 “The goal is to provide an accurate picture of the experience associated with activities . . . .”20 The DRM seeks to mimic, at much lower cost and disruptiveness, methods based on “experience sampling,” in which well-being is measured by asking respondents, a number of times throughout the day, to “record where they are, what they are doing, and how they feel”21 – though (as elaborated below) the DRM is quite nuanced in how it assesses people’s feelings. By evoking the context of the previous day, the DRM seeks to “elicit specific and recent memories, thereby reducing errors and biases of recall” even though perceived well-being is not recorded during the activity, as it is in experience sampling.22 Moreover, while experience sampling provides “a sampling of moments,” the DRM provides “an assessment of contiguous episodes over a full day.”23 Well-being in the DRM is based on respondents’ ratings, from a low of zero to a high of six, on a long list of positive and negative affect measures (e.g., angry/hostile, worried/anxious, happy, warm/friendly, enjoying myself, frustrated/annoyed, depressed/blue, hassled/pushed around).24 Patterns on the

17 See, e.g., Daniel Kahneman & Alan B. Krueger, Developments in the Measurement of Subjective Well-Being, 20 J. ECON. PERSP., Winter 2006, at 3, 6. 18 See generally http://www.krueger.princeton.edu/subjective.htm, supra note 9 (select “Tx DRM Stata Data File”); World Values Survey, http://www.worldvaluessurvey.org (last visited Mar. 9, 2010); General Social Survey, http://www.norc.org/GSS+Website (last visited Mar. 9, 2010). 19 Daniel Kahneman, Alan B. Krueger, David A. Schkade, Norbert Schwartz & Arthur V. Stone, A Survey Method for Characterizing Daily Life Experience: The Day Reconstruction Method, 306 SCI. 1776, 1776 (2004). 20 Id. 21 Id. 22 Id. at 1776-77. 23 Id. at 1777. 24 Id. at 1777 tbl.1.

2010] “LIVING WELL” AND WELL-BEING 645

DRM prove to be similar in many respects to those obtained with the much more costly experience sampling method.25 Table 1 shows average ratings for the four most positive and four least positive activity categories in the Princeton group’s DRM data. In evaluating people’s perceived well-being, the DRM calls attention not to the global assessments of life satisfaction tapped by the World Values Survey and General Social Survey questions but, instead, to perceived well-being while performing specific life activities such as those in Table 1 – socializing, working, commuting, etc.

Table 1: Most Positive and Least Positive Activities in the DRM Data26

Mean Mean Proportion Mean positive negative of sample hours/day affect rating affect rating reporting Most positive activities Intimate relations 5.10 0.36 0.2 0.11 Socializing 4.59 0.57 2.3 0.65 Relaxing 4.42 0.51 2.2 0.77 Pray/worship/meditate 4.35 0.59 0.4 0.23 Least positive activities Computer/e-mail/internet 3.81 0.80 1.9 0.47 Housework 3.73 0.77 1.1 0.49 Working 3.62 0.97 6.9 1.00 Commuting 3.45 0.89 1.6 0.87

What are we to make of the focus on the “performance” of life’s activities, rather than the “product” or overall narrative of a life, in Dworkin’s conception, and in the Princeton group’s data?

II. GENUINENESS In considering these otherwise divergent approaches − Dworkin’s and the Princeton group’s − it is useful to return to Dworkin’s assertion that “[w]e value human lives as these are lived not for the completed narrative, as if fiction would do as well, but because they too embody a performance: rising to the challenge of having a life to lead.”27 I will be suggesting that one thing this

25 Id. at 1777-78. In addition to the DRM, the Princeton group has conducted the Princeton Affect and Time Survey (“PATS”). See Krueger et al., supra note 8, at 34-36. The PATS data are even less burdensome to collect than the DRM data; however, they are also much less closely correlated with experience sampling results. See id. at 39 fig.1.5. I focus on the DRM, for which, as noted, the full dataset is publicly available. 26 The data appears in Kahneman et al., supra note 19, at 1777 tbl.1. All individuals in the sample were employed. For further details on the sample, see id. at 1777. 27 DWORKIN, supra note 2 (manuscript at 124) (emphasis added).

646 BOSTON UNIVERSITY LAW REVIEW [Vol. 90:641 passage − particularly when viewed in light of the empirical literature on self- reported well-being − does is surface the issue of genuineness in perceived well-being. More particularly, I will be suggesting that overarching life satisfaction questions, such as those on the World Values Survey and the General Social Survey, create the risk that the answers given will not be “genuine,” as those answers will often be not actual evaluations made apart from the posing of the survey question but, instead, answers – narratives – that are crafted in one’s mind in response to artificial contemplation of the survey question. Most of us have a reasonably clear sense of whether – to use the DRM descriptors – we are “happy,” “anxious,” “hassled,” and so forth during the performance of a typical activity over the course of a given day – and we have this sense regardless of whether we are asked by a survey instrument about our experience – but the same cannot necessarily be said about the World Values Survey and General Social Survey questions. Most of us do not always walk around with a clear answer to that sort of question in our heads. Thus, an “adverbial” focus on the performance of life’s activities, a more direct and (for many individuals) more natural object of evaluation, will tend to reduce the prospect of reports that, rather than experienced, are constructed in response to a survey’s posing of a particular question. The analytic perspective here differs from that of the Princeton group, which tends to view responses to global life satisfaction questions not so much as lacking “genuineness” but, instead, as a form of cognitive error. In the words of the Princeton group, thinking about global life satisfaction “may induce a form of focusing illusion, by drawing people’s attention to their relative standing” in (primarily material) life circumstances.28 On this account, the focusing illusion distorting people’s answers is likely to arise even apart from the prompt of a World Values Survey or General Social Survey question; individuals who, unprompted by a survey, assess their overall life satisfaction will nonetheless, on the Princeton group’s view, tend to suffer from cognitive failure in the form of a focusing illusion. The alternate perspective in this Essay is that – building from the suggestion of the Dworkin passage above – the posing of overarching life satisfaction questions on an instrument such as the World Values Survey or the General Social Survey may lead to an artificially crafted answer that lacks “genuineness” (even if it cannot be termed an “error”). I will not be making a normative claim in favor of “genuineness”; conceivably, one could view well-being assessments in response to a survey’s overarching life satisfaction question as “better” in some respect or another than DRM assessments even assuming the latter assessments are more “genuine.” While I will not be making a general normative claim, Parts III and

28 See Daniel Kahneman, Alan B. Krueger, David Schkade, Norbert Schwarz & Arthur A. Stone, Would You Be Happier If You Were Richer? A Focusing Illusion, 312 SCI. 1908, 1908 (2006).

2010] “LIVING WELL” AND WELL-BEING 647

IV will offer some data-driven suggestions of the possibly appealing features of “genuineness” in the contexts of income and sociability effects on self- reported well-being. Note that “genuineness” does not in any way preclude a role for an overarching life satisfaction perspective in a given individual’s life (and indeed it is not easy to imagine a life in which no such overarching perspective is present in any domain). “Genuineness” here is concerned with the construction of an answer because a survey posed a question, not on the range of questions some individuals may choose to ask themselves as they go about living their own lives. Just as, for a discrete episode within daily life, experienced well-being is unlikely to be, to a very large degree, an artifact of measurement, for some individuals an assessment of overarching life satisfaction may preexist – rather than be constructed in response to – a question on the World Values Survey or the General Social Survey, and both types of well-being answers will be “genuine” for such individuals. As much of the discussion below will reveal, however, the Princeton group’s data suggest at least some possibility that many individuals may be closer to Amartya Sen’s “Dora” in this Symposium – a figure for whom it is “silly” to ask “whether her life is successful (‘there are so many more interesting things to think about’).”29 As an opening lens onto the issue of “genuineness” I am seeking to explore here, consider what we know about the role of markers of material success − such as a large and impressive home − on individuals’ self-reported well-being in the existing empirical literature. As briefly noted above, when people are asked about their global life satisfaction on the World Values Survey and the General Social Survey, people’s broad-scale image of themselves, their accomplishments, and their life seems to figure significantly. One important aspect of this image, at least in America, will often be their house; if that house is a large, impressive structure on lots of land in a leafy suburb, then it appears to have a significant positive effect on self-reported global satisfaction, even after controlling for income.30 One of the clearest insights to emerge from DRM measures of experienced well-being, however, is that isolation from others, and a long car commute to and from work, greatly reduce well-being as reflected in these measures. Table 1 above shows that commuting finishes dead last in people’s ranking of activities, and the evidence on sociability in Part IV below is equally unequivocal.31 The likely contrast between the global satisfaction perspective and the performance-of-activities perspective is

29 See Amartya Sen, Dworkin on Ethics and Freewill, 90 B.U. L. REV. 657, 659 (2010). 30 See Grace Wong, Be It Ever So Humble: Understanding Housing Using Subjective Well-Being Data 18-19 & tbl.12 (Feb. 2007) (unpublished manuscript), available at http://real.wharton.upenn.edu/%7Ewongg/research/swb.pdf. 31 The Wong study does not directly investigate the relationship between home value and DRM measures, though she finds that home value is largely uncorrelated with the gap between DRM “affect” at home versus away from home, see id. at 17 & tbl.9.

648 BOSTON UNIVERSITY LAW REVIEW [Vol. 90:641 wonderfully encapsulated in the rueful comment of behavioral economist George Loewenstein about his own purchase of a large, distant home: We often yearn for a roomy, isolated home . . . when, in fact, it will probably compromise our happiness by distancing us from neighbors. (Social interaction and friendships have been shown to give lasting pleasure.) The big isolated home is what Loewenstein, 48, himself bought. “I fell into a trap I never should have fallen into,” [Loewenstein] told me.32 If a large outlying home has divergent effects on perceived well-being depending on the measure used, it is clear that the DRM measure is the one likely to be more “genuine.” Repeated diary reports of misery during long car commutes seem likely to reflect an actual experience of dissatisfaction rather than an answer artificially constructed in response to the asking of a well-being question. Unless many of those who buy homes like Loewenstein’s spend meaningful amounts of time on a typical day in abstract contemplation of their roomy mansion – and yet somehow the effects of such contemplation are not picked up by the DRM measure − the DRM measure will tend to be the more “genuine” one. The remark just made highlights an important gap in the existing empirical well-being literature, which has not yet addressed the strength of the link between (“genuine”) overarching assessments actually made in the course of people’s lives (rather than in response to being asked a question on a survey) and people’s measured well-being on the DRM. Presumably because the Princeton group’s analytic views overarching satisfaction questions as creating the risk of cognitive error (as described above), the Princeton group has thus far not explored how global satisfaction assessments people actually make in the course of their lives may affect their DRM evaluations – yet surely the relationship may be an important one in at least some contexts. When, for instance, a scholar painstakingly reviews the source citations in an article prior to its publication, the task would presumably be far more unpleasant (if not completely unbearable) were the scholar not to conceptualize the task as one component of a general practice of carefully preparing scholarly work for publication. Or, to take a different example, a Feminine Mystique acolyte would presumably view people who think they are gaining deep fulfillment from waxing their kitchen floors as basing their evaluations on a form of ideology that had filtered into their DRM assessments of the cleaning task; as Betty Friedan wrote nearly fifty years ago: [The] mystique of feminine fulfillment became the cherished and self- perpetuating core of contemporary American culture. Millions of women lived their lives in the image of those pretty pictures of the American suburban housewife, kissing their husbands goodbye in front of the

32 Jon Gertner, The Futile Pursuit of Happiness, N.Y. TIMES, Sept. 7, 2003, § 6 (Magazine), at 44, 86.

2010] “LIVING WELL” AND WELL-BEING 649

picture window, depositing their stationwagonsful of children at school, and smiling as they ran the new electric waxer over the spotless kitchen floor. They baked their own bread, sewed their own and their children’s clothes, kept their new washing machines and dryers running all day.33 Undoubtedly, the performance of many acts cannot be viewed in isolation from broader themes (of varying normative valence) that run through and give structure to our lives. For some, the experience of a profession or occupation – like Loewenstein’s experience of his large outlying home – may diverge depending on the perspective on well-being taken, the global assessment or the day-to-day measure of the experience of work tasks and activities. In Kahneman and Krueger’s view, “Respondents who answer abstract evaluative questions about [their work] are likely to be reminded that [work is a desirable aspect] of their life. Reports of how much they enjoy [working] will tend to be anchored on that general assessment, resulting in a favorable bias” on well-being from work compared to respondents’ DRM assessments of work.34 For Kahneman and Krueger, “[B]iases are reduced in the Day Reconstruction Method, in which respondents describe particular episodes, without reference to more general evaluations of parts of their lives.”35 Again, even if the overarching assessment is not a form of error, it is highly plausible that the more “genuine” assessments are those captured by the DRM. Certain occupations provide a clear illustration of the earlier suggestion of the normative complexity of “genuineness”; medical professionals, for instance, who treat patients with small cell lung cancer – a rapid and ruthless form of cancer that often quickly takes the lives of those (many not yet even in middle age) who contract it36 – might not offer many positive responses on the DRM to their lived experiences at work. But if indeed they are not obtaining “genuine” well-being from their work, a negative evaluation obviously does not follow. To return to Dworkin’s product-performance distinction from above, perhaps it is hard to define such a medical professional’s “performance” of life without strong reference to the “product” of such a life.37 Similarly, in the domain of self-reported well-being, a positive response by such a medical professional to a World Values Survey or General Social Survey global life satisfaction question might be normatively preferable to a DRM response, even if it is less “genuine.”

33 BETTY FRIEDAN, THE FEMININE MYSTIQUE 18 (1963). 34 Kahneman & Krueger, supra note 17, at 13. 35 Id. 36 See, e.g., Noni – a 33 Year-Old Victim of R.J. Reynolds’ Camel Cigarettes, http://whyquit.com/whyquit/A_Noni.html (last visited Mar. 9, 2010). 37 Cf. George Loewenstein, That Which Makes Life Worthwhile, in MEASURING THE SUBJECTIVE WELL-BEING OF NATIONS, supra note 8, at 87, 97-98 (describing Loewenstein’s wife’s experience of a holiday spent caring for an ill parent, and questioning whether those caregiving tasks increased her perceived well-being in a DRM sense).

650 BOSTON UNIVERSITY LAW REVIEW [Vol. 90:641

If there are times, such as these, in which “genuineness” may not be normatively compelling, the next two Parts offer data-driven accounts in which “genuineness” seems to hold up far better in a normative calculus.

III. SELF-REPORTED WELL-BEING, INCOME, AND GENUINENESS In this Part and the next, I consider two basic questions about self-reported well-being. The first is: “What is the role of income in perceived well-being?” Beginning with the global life satisfaction measures, an exhaustively discussed feature of these measures is their relatively modest correlation with income. To some degree, when our income rises, we adapt to the higher income level and do not give greatly more positive answers to questions such as “All things considered, how satisfied are you with your life as a whole these days” or “Taken all together, how would you say things are these days?”38 Such results may provide an empirical counterpart to Robert F. Kennedy’s famous suggestion that the focus of policymakers on the Gross National Product means that they fail to capture “the strength of our marriages . . . our courage . . . our wisdom . . . our learning . . . our compassion . . . in short, [everything that] makes life worthwhile.”39 While the correlation between global life satisfaction measures and income is thus more modest than some might expect, a sensible prediction is that the correlation of DRM well-being measures with income will be much lower. As the Princeton group suggests, “[t]he task of evaluating one’s life” in response to the overall life satisfaction question “evokes a comparison of one’s objective circumstances to conventional standards, reminding the rich that they are rich,” while the “relative impact of [income] on actual affective experience” is likely to be much smaller.40 Thoughts of one’s income “are much more likely to come to mind when answering questions about one’s life than in the routine course of experience.”41 In other words, whatever the role of income in answering a question about one’s satisfaction with (as the World Values Survey puts it) “life as a whole,” income’s role will be much more limited in the assessment of the enjoyment of performing activities such as “socializing” or “praying” from Table 1 above. In fact, the Princeton group’s data show exactly this pattern. Income correlates four times as much with life satisfaction (r=0.20) as with net affect on the DRM (r=0.05).42 (“Net affect” is the gap between positive and negative

38 See generally, e.g., Richard A. Easterlin, Will Raising the Incomes of All Increase the Happiness of All?, 27 J. ECON. BEHAV. & ORG. 35 (1995). 39 See Senator Robert F. Kennedy, Remarks at the University of Kansas (Mar. 18, 1968) (transcript available at http://www.jfklibrary.org/Historical+Resources/Archives/Reference+ Desk/Speeches/RFK/RFKSpeech68Mar18UKansas.htm). 40 Kahneman et al., supra note 19, at 1779. 41 Id. 42 See id.

2010] “LIVING WELL” AND WELL-BEING 651 affect, the two affect measures reported in Table 1.43) And the Princeton group finds the same pattern in comparing the effects of income on overall life satisfaction versus well-being measured through experience sampling.44 In fact, in the Cornell Work Site Blood Pressure Study data that they examine in addition to their own dataset, the correlation between income and the average “happiness rating” during the day was a mere 0.01.45 While this Essay does not defend the value of “genuineness” across the board, with respect to the effect of income on self-reported well-being, the more “genuine” DRM measure seems normatively appealing in yielding the reassuring conclusion that a high income is spectacularly unimportant to our well-being. (Note that the data here come from advanced nations; in a nation characterized by extreme poverty, both the data and the analysis would almost certainly be different.46) At least in this domain, “genuineness” picks out a well-being measure that seems normatively attractive on other grounds.

IV. SELF-REPORTED WELL-BEING, SOCIABILITY, AND GENUINENESS Turning from the effects of income to the effects of sociability, we discover in the alternative empirical measures of perceived well-being exactly the opposite of what occurs with respect to income; global life satisfaction is only modestly affected by sociability, while DRM measures are very much affected by it. We saw above that the correlation of global life satisfaction with income was approximately four times that of the DRM net affect measure with income.47 While the Princeton group’s published work does not report parallel global life satisfaction and net affect results related to sociability, we can generate comparable statistics using the same methodology and employing the Princeton group’s publicly available dataset.48

43 See id. 44 Kahneman et al., supra note 28, at 1909. 45 Id. 46 See, e.g., Easterlin, supra note 38, at 40. 47 See supra note 42 and accompanying text. 48 See http://www.krueger.princeton.edu/subjective.htm, supra note 9 (select “Tx DRM Stata Data File”). As a check on my use of the dataset, I first sought to replicate the Princeton group’s published correlations of income with global life satisfaction and net affect. Global life satisfaction in the Princeton group’s dataset is measured by “q1,” which takes values from “very satisfied,” translated numerically as 1, to “not at all satisfied,” translated numerically as 4. (The translations carry the potential for confusion because someone who is more satisfied has a lower numerical value of q1.) After dropping observations that duplicate q1 and the log of household income (which are the same for each observation corresponding to a given respondent in the Princeton group’s dataset), the correlation of q1 with the log of household income (“lghsinc” in the dataset) is -0.20 (N=907). Because global life satisfaction is higher when q1 is lower (as just noted), the correlation of global life satisfaction with the log of household income is 0.20, as reported by the Princeton group, see supra note 42 and accompanying text.

652 BOSTON UNIVERSITY LAW REVIEW [Vol. 90:641

Within the variables contained in the Princeton group’s dataset, sociability is best measured by the proportion of daily episodes spent with others. (One question on the Princeton group’s survey is whether the respondent was interacting with anyone during a given diary episode.49) The correlation of global life satisfaction with this proportion turns out to be barely one-fiftieth the size of the correlation of DRM net affect with this proportion (r=0.002 versus r=0.089).50 While much of the empirical literature on self-reported well-being as measured by either approach has rightly emphasized the general importance of sociability, this factor turns out to be of far greater importance for DRM well-being than when people are answering a question such as the World Values Survey’s query about “life as a whole.”51 Further evidence of the importance of sociability to achieving well-being from the actual performance of life’s activities is provided by the Princeton group’s comparison of respondents’ answers to a general question about the enjoyability of “socializing after work” and actual DRM ratings of such socializing, which were much higher than ratings in response to the general

I was also reasonably successful in my effort to replicate the Princeton group’s 0.05 correlation between DRM net affect and lghsinc, though the correlation I obtained, 0.057, is slightly different from the Princeton group’s figure. To obtain the 0.057 figure, I first generated an average net affect value for each individual respondent’s identification number (“id” in the Princeton group’s dataset) by summing the net affect values (“netaff” in the dataset) for each id (in Stata, “by id: egen t_netaff=total(netaff)”) and then dividing by the number of observations for that id (“by id: gen avg_netaff=t_netaff/(adjusted + adjcode)” in Stata, where “adjusted” and “adjcode” relate to the number of observations for a given individual in the Princeton group’s dataset). After dropping observations that duplicate avg_netaff and lghsinc (which, again, are the same for each observation corresponding to a given respondent), the correlation of avg_netaff with lghsinc is 0.057 (N=908). 49 See http://www.krueger.princeton.edu/subjective.htm, supra note 9 (select “Documentation & Questionnaire”), at 20, 22, 24, 26, 28, 30, 32, 34, 36, 38, 40, 42. 50 To obtain the correlations, I first generated a dummy variable that took on a value of 1 if the respondent was interacting with another person during a given episode, using the “alone” variable in the Princeton group’s data (Stata command “gen nalone=1 if alone==0”). (Examining whether the follow-on question about the identity of interaction partners for those who were interacting with someone during a given episode was asked confirms the natural expectation that “alone” in the Princeton group’s data takes on a value of 0, rather than 1, if the respondent was interacting with another person during the episode.) Next I generated the proportion of daily episodes spent with others for each individual identification number (“by id: egen t_nalone=total(nalone)” and then “by id: gen f_nalone=t_nalone/(adjusted + adjcode)” in Stata). The correlation of q1 (the global life satisfaction variable) with the proportion of daily episodes spent with others was -0.002 (N=908), implying a correlation of 0.002 between global life satisfaction and that proportion. The correlation of avg_netaff, see supra note 48, with the proportion of daily episodes spent with others was 0.089 (N=909). 51 For a discussion of the overall importance of sociability to self-reported well-being, including reference to many other studies, see JONATHAN HAIDT, THE HAPPINESS HYPOTHESIS: FINDING MODERN TRUTH IN ANCIENT WISDOM 133 (2006).

2010] “LIVING WELL” AND WELL-BEING 653 question.52 It should be noted, however, that while the subjects who offered the two types of ratings were selected in the same manner,53 those who actually engaged in “socializing after work,” and thus reported ratings of such socializing in the DRM context, may not represent a random subset of the overall sample.54 In other words, it is possible that differences in the two types of ratings reflect, at least in part, differences in the samples of respondents providing ratings (depending on whether they actually socialized after work) rather than differences in views of socializing in the abstract versus actual performance of socializing activities. The normative thrust of sociability’s effects may be somewhat less clear than with respect to income, but to the extent a significant role for sociability in experienced well-being is viewed as normatively desirable, the more “genuine” DRM measure again turns up as normatively appealing.

V. LIVING WELL? The unexpected, though unquestionably only partial, mapping described here between Dworkin’s product-performance distinction and the two alternate empirical approaches to self-reported well-being is offered in the spirit of a modest illumination from interchange of ideas. Of course, the two paradigms differ greatly in their commitments and premises, and this Essay does not seek to minimize those differences. Notwithstanding the differences, the discussion above of the question of how overarching life satisfaction assessments relate to DRM measures of lived daily experience serves as an interesting pointer to a potential complexity in “living well” for Dworkin. In his account, the priority of “performance” over “product” in “living well” is followed by an identification of the importance of both “self-respect” and “authenticity” as components of “living well.”55 “Self- respect” requires that each person accept that successful performance of life is a matter of importance, while “authenticity” requires that each person take “responsibility to create [a] life through a coherent narrative that he himself has chosen and endorses.”56 What is the relationship between the emphasis on choosing a narrative to achieve “authenticity” and, thus, “live well,” on the one hand, and Dworkin’s earlier dismissal of “the completed narrative” in lieu of the act of performance, on the other? Dworkin does not develop the relationship, and while – as noted earlier in this Essay – at least some conception of an overarching life narrative is of potential importance to even a DRM-oriented conception of perceived well-being,57 as well as of course

52 See Krueger et al., supra note 8, at 35 tbl.1.2. 53 See id. at 32. 54 See id. at 33 (“With . . . some possible differential selection as to who participated in the activities on the diary day, the results should be read cautiously.”). 55 DWORKIN, supra note 2 (manuscript at 128). 56 Id. (emphasis added). 57 See supra text accompanying note 33.

654 BOSTON UNIVERSITY LAW REVIEW [Vol. 90:641

(through “authenticity”) to “living well,” perhaps the precise nature and attributes of narrative’s role in “living well” would profit from further analysis by Dworkin.

CONCLUSION As emphasized above, the concept of “genuineness” emphasized in this Essay is not given a general normative defense. Moreover, some form of “constructed” answer – even if “genuine” rather than prompted by a survey question – may be simply unavoidable in circumstances in which the assessment and the thing to be assessed are significantly temporally removed. The Princeton group’s position on this issue is clear: “[A]ffective experiences are fleeting and not available to introspection once the feeling dissipate[s].”58 It is certainly possible that the Princeton group overstates the point for certain common feelings or experiences (for example, waking up tired after staying up too late); we may be able to access those feelings quite directly, without a serious prospect of any form of construction of a response, even when we are not close temporally to experiencing them. Alternatively, the Princeton group may indeed be right in its suggestion that (in the words of the fictional Costals quoted at the outset – a character I quote with some reluctance given how unappealing he is) “[h]appiness writes white.”59 Perhaps nothing other than a significant act of construction is possible insofar as temporally distant experience is concerned. Even without either temporal distance from a discrete event or the prompt of a general life satisfaction question on the World Values Survey or the General Social Survey, it is of course true that some individuals may engage in forms of motivated fictionalizing in making many different forms of life assessment. While the medical example above underlines the notion, widely held among scholars, that in some cases removal from actual lived experience will be normatively attractive, the downsides of such dissociation are perhaps less recognized. Of course, some examples may be easy; many of us know someone whose self-perception seems to reflect an astonishing failure to confront basic features of lived reality and, one guesses, the level of experienced well-being that would be reflected if an assessment were made in accordance with either the Princeton group’s DRM measure or the following beautiful passage from Dworkin: When you do something smaller well – play a tune or a part or a hand, throw a curve or a compl[i]ment, make a chair or a sonnet or love – your satisfaction is complete in itself. It needs no bush and it needs no sequel.

58 Krueger et al., supra note 8, at 29; cf. ARISTOTLE, supra note 4, bk. X, at 259 (“[T]he pleasures involved in activities are more proper to them than the desires; for the latter are separated both in time and in nature, while the former are close to the activities . . . .”). 59 DE MONTHERLANT, supra note 1, at 157. For discussion of Costals’s misogyny, see Ordeal by Hippogriff, TIME, Feb. 7, 1969, at 98 (book review), available at http://www.time.com/time/printout/0,8816,838962,00.html.

2010] “LIVING WELL” AND WELL-BEING 655

Those are achievements within life. Why can’t a life also be an achievement complete in itself with its own value for the art in living it displays?60 In other words, in some cases it is hard to imagine high marks being given to the life actually being lived − and yet a more positive response to an overall “life narrative” question seems far more imaginable because of the greater ease of fictionalizing. Sadly, the decisions of such individuals may have large negative effects on others. This observation in turn highlights a further fascinating feature of performance-oriented accounts of well-being: such accounts seem more connected to others’ lived experience. While people’s internally constructed narratives of their lives may bear little or no relationship to the views and the experiences of others, a person’s daily lived experience − precisely because it is less easily fictionalized − seems unlikely to be as divorced from others’ well-being. Of course, this point raises a number of complexities that − like much else in Dworkin’s sweeping book − cannot be explored fully in this Essay.

60 See DWORKIN, supra note 2 (manuscript at 125).